It should also be up to the developers of apps. Some may not care (ex: beat saber), or someone just want an ability to ban from their app specifically for bad behavior in their app, or have cool off periods instead of perma bans or whatever else.
One potentially error-prone part of moderating any site with community content is handling users evading bans by creating new accounts. If you don't have any mechanism to prevent that, you have a much larger amount of offensive or troll posts, and react slower because you don't have the history of the user because you treat each account as a different person. But any mechanism trying to detect if a new user is potentially a sock puppet of a troublemaker is acting on rather few bits of information. So false positives are not that unlikely.
(A ban from multiplayer modes or just voice chat would be more appropriate, perhaps?)